Some data processing applications, such as computer-aided design (CAD) and similar graphics programs, generate large volumes of data. CAD programs, in particular, produce such cumbersome data files that least a portion of the data is preferably translated into a “lightweight” format for viewing, markup, mockup, styling and analysis operations. For instance, the DirectModel family of CAD applications available from Electronic Data. Systems Corporation of Plano, Tex. employs “JT” files (for the former “Jupiter Toolkit” name, using the file extension) that contain several types of data critical for digital mockup operations, including: three dimensional (3D) geometric representations of CAD parts and assemblages thereof, used for high-speed interactive visualization, picking, approximate measurement, and approximate collision and clash detection; precise boundary representation (Prep) of each CAD part, used for precise measurement and precise collision and clash detection; product manufacturing information (PMI) exported from the original CAD source; and additional textual meta-data exported from the original CAD source.
Enterprises employing such translated lightweight CAD files typically have significant economic resources invested in the processes of producing, coordinating and warehousing the data, along with the substantial physical resources employed, to actually store the data. The demands for storage space increase as, for instance, the automobile and aerospace industries produce more lightweight data and construct ever-larger and/or higher-fidelity assemblies. Therefore making the lightweight files as small as possible, by any practical means available, is of primary concern.
At the same time, however, demands for visualization performance (i.e., image data retrieval and rendering). Streaming, in this context, refers to the ability to continuously load 3D geometric model data without forcing the user to wait. The user can meaningfully interact with a model while the data is being loaded, and without suffering any perceivable performance degradation. Reducing the size of lightweight data files should retain the option of retrieving the contents a piece at a time.
Subject to that constraint, every link in the application chain leading to efficient large-model analysis and visualization wins when the lightweight files become smaller: files become smaller (and fewer disk drives or similar storage media need be purchased), disk access time decreases, network utilization is decreased, and the perceived level of application performance is increased due to quicker loading of model data.
There is, therefore, a need in the art for a system, process and data format providing more compact lightweight CAD data files without substantial degradation of visualization performance.